The biological actions of estrogen are manifest through two genetically distinct estrogen receptors (ER alpha and ER beta) that display nonidentical expression patterns in target tissues. The phenotypic alterations in response to estrogens in mice disrupted for either or both of these receptors are not identical, suggesting that each subtype plays a unique role in ER-action. However, the lack of subtype-specific agonists and antagonists has made it difficult to define the processes that are regulated by ER alpha and/or ER beta. Previously, we have reported the identification and characterization of a series of LXXLL-containing peptide antagonists that block estrogen signaling by preventing the association of ER alpha with required coactivators. As expected, given the similarity of the coactivator binding pockets among nuclear receptors, most of the peptide antagonists identified inhibited the activity of multiple receptors. However, by altering sequences flanking the core LXXLL motif, some receptor selectivity was afforded. Building on this observation, we have screened combinatorial phage libraries, expressing peptides in the format X7LXXLLX7, for peptides that interact in a specific manner with ER beta. Using this approach, a series of highly specific, potent peptide antagonists have been identified that efficiently inhibit ER beta-mediated estrogen signaling when introduced into target cells. Interestingly, in cells where both ER subtypes were expressed, these ER beta antagonists were capable of attenuating ER action, suggesting that ER alpha and ER beta do indeed form functional heterodimeric complexes. We believe that suitably formulated versions of these peptides can be used to study ER beta action in vitro and in vivo. In addition, the unanticipated specificity of the peptides identified should serve as an impetus to investigate the use of this approach to develop peptide antagonists of other nuclear receptors and unrelated transcription factors.